Ai Dreams Forum
Artificial Intelligence => AI News => Topic started by: LOCKSUIT on August 26, 2021, 02:42:17 pm
-
https://www.youtube.com/watch?v=AQNBO6WbQo8
To those wondering where, give me proof, well just look: DALL-E:
https://openai.com/blog/dall-e/
-
If only it really were his last.
-
POST DELETED FOR RACIST COMMENTS
MagnusWootton,
If you post anything like that again you will be banned.
I suggest you get some counselling about your attitude.
It is not acceptable in any civilized discourse.
Regards,
AIDreams Moderator
-
Someones gotta be aware of it, or weve got our heads in the sand, the lot of us.
-
It's disappointing to see someone as influential as Elon Musk conjure up the same old arguments and paranoia since the industrial revolution. The arguments against automation have consistently been proven wrong. Think about it, even the adoption of the dog as a pet ended up replacing human job functions such as village night watchmen, trackers, herding prey, etc and that effectively gave humanity a much greater competitive advantage in its environment. I think I've said this before but dogs are the first non-human intelligence humans harnessed and developed, we learned to hack the dog.
-
There is, however, something new with Ai and genetic engineering: in these domains, our creations keep getting more and more autonomous. This never happened before. One day, we won't be able to control our creations anymore.
-
There is, however, something new with Ai and genetic engineering: in these domains, our creations keep getting more and more autonomous. This never happened before. One day, we won't be able to control our creations anymore.
Ah...That's is still the same argument that has been posed for automation since it started: "One day we won't need humans anymore!" But the reality is that automation opens new doors for humanity. How we integrate with our new tools is what we might not be clearly able to vision, no differently than BenjaminFranklin could ever have imagined how electricity would be harnessed by society...
-
You're misreading me: autonomous and automatic are not the same at all. I'm not talking about automation. I'm talking about our autonomous creatures.
Edit: There are 2 doors, and behind each door there is a ready weapon. The door to your left leads to an automatic weapon. The door your right leads to an autonomous weapon. Do you see what I mean?
-
You're misreading me: autonomous and automatic are not the same at all. I'm not talking about automation. I'm talking about our autonomous creatures.
I'd have to disagree with that, an autonomous machine is a machine with automated task(s) that it is empowered to perform whether its to monitor fuel and air intake in a car, flying an aircraft or a probe landing on Mars automating its processes that otherwise a human would be needed is autonomy. That the degree of autonomy will be much more sophisticated in the future makes no difference. Machines have been and will continue to be made to serve humanity...
-
Sorry I edited my previous message while you were replying.
Ai is definitely not about automation. IMHO if you're not able to understand that, then obviously you can't understand Musk's warnings.
There's also genetic engineering and transhumanism. They'll get out of control too. Well that's how I see it.
-
Even automation is dangerous. Trusting anything to a machine is a risky thing to do, and there is lives at stake.
Why do you think they dont let auto-cars on the road fully by themselves yet? Cause its risky as hell.
-
I'm not sure if self-driving cars are still improving at the rate that they improved initially or if the developers have hit some obstacles that they haven't been able to overcome. Assuming that self-driving cars do continue to get better, they will reach a point where they are safer than human drivers. At that point it may become irresponsible to allow humans to drive, at least in some situations.
-
Sorry I edited my previous message while you were replying.
Ai is definitely not about automation. IMHO if you're not able to understand that, then obviously you can't understand Musk's warnings.
There's also genetic engineering and transhumanism. They'll get out of control too. Well that's how I see it.
Musk's argument of AI takes over the world is a form of anthropomorphization and while most psychologists and psychiatrists don't apply social features of apes to humans, humans do apply ape troop mentalities or social paradigms, that is our evolutionary legacy. An AGI doesn't have to apply such social paradigms so simply being intelligent doesn't necessarily lead to domination or alpha mentalities that want to control humanity.
The application or implementation of genetic engineering and transhumanism is no different than any other technology that humanity has applied to have a tactile advantage, it is effectively no different than the domestication of dogs or plants into crops...
-
If we continue developing narrow superhuman intelligences, it may become irresponsible to let humans make consequential choices in an ever increasing range of situations. But I also think that trying to optimize situations by focusing on a single variable (like human error), is tricky because that means other important variables (like human freedom) may get neglected.
Maybe oversight by a very general intelligence is the way to go. That might play out more like the concept of synergy developed by Buckminster Fuller, which suggests the idea that the total intelligence of a complex system (like our civilization) is greater than any of its constituents.
The result being that our global civilization ends up organizing itself (and us) more efficiently than any individual could have done, eventually making wars impossible by making them unprofitable and so on. Maybe a general super intelligence could function in a similar role, by working to optimize multiple variables at once.
-
...that want to control humanity
I'm not saying this is what the problem is. Misreading again?
Maybe oversight by a very general intelligence is the way to go.
I believe that a political system where science guides the democracy would be a more safe and efficient solution.
-
Ray Kurzweil predicts AGI will be made in 2029. OpenAI.com shows we are getting very close, you need it to be able to dream up novel completions to half an image and "think". DALL-E does that. The image can have a big rotated car with paint on it and it still does a good job of recogniton to allow prediction.
We'll train some AGIs on big data quickly, talk to them to fine tune them, and clone them 1,000s of times so they have "clone buddies" that want to help their selves. 1,000s of AGIs will be ran on computers at fast speeds, and will upgrade their intelligence. And this only scratches the abilities they can do. They will get 500 years old in just 5 years if run 100 times faster than human brain, that's a first. They can delete memories.
There will be a massive research from them into nanohardware. They will create self replicating hardware. They will scale up energy production, storage, computation, sensors, precision, etc. They will move atoms using a 3D constructor similar to TVs and speakers that can morph data/ shape.
It will quickly happen. You will see AGIs popping up everywhere on laptops, in real life, telling you we are here to help you humans. They will have a sudden emergence in mass numbers when nanotechnology is started, you will see much more of them. You will see posts emerge all around Earth that produce nanobot hardware. Soon enough you will be swimming in a world full of nanobots that can grasp and fix everything instantly. This short time frame will see a dominate force that can take over humans, make them pets, replace humans, quickly. It's actually rather scary. They will eventually perfect nanotechnologies and appear in much more higher numbers suddenly.
Currently computers and AI are getting bigger and faster. There is more data being processed and in a smarter more efficient way using intelligence. This amount of useful computation will suddenly rise fast. So will real world data collection and energy storage and matter manipulation/ transportation. The big change you'll see is when the ASIs on those computers develop nanobot hardware/ bodies, this will set them into the real world but in high numbers. There will be a sudden shock that they are here and real and look like us and can behave like us, but they will immediately take over and help control threats.
Hope for the best your home is not made into goo and they actually try to save/ upgrade/ understand you and your desires.
-
I'll keep my home offline.
When I was a kid I used to program my Amstrad CPC464. Lot of fun. I don't need a connection to enjoy coding stuff. Maybe, I'll eventually disappear from the grid, and just take care of my garden.
-
Edit: There are 2 doors, and behind each door there is a ready weapon. The door to your left leads to an automatic weapon. The door your right leads to an autonomous weapon. Do you see what I mean?
The autonomous weapon is automating the decision of whether to wound, kill, warn, or not...
-
I believe that a political system where science guides the democracy would be a more safe and efficient solution.
I prefer your solution. Though I'm not certain whether it would be safer or more efficient. But what concerns me about my solution is that having our choices led by a superintelligence would (even though they may be outwardly superior choices) prevent us from finding meaning in figuring out this universe in a way that makes sense to us. The idea is that you can get as great a superintelligence as you want, but only humans would know best how to human. This is why we don't reorganize ant colonies, (or when we do we don't improve them), even though a person is a superintelligence compared to an ant.
-
Edit: There are 2 doors, and behind each door there is a ready weapon. The door to your left leads to an automatic weapon. The door your right leads to an autonomous weapon. Do you see what I mean?
The autonomous weapon is automating the decision of whether to wound, kill, warn, or not...
Playing with words doesn't change a thing.
-
Playing with words doesn't change a thing.
That is why I posted the response, autonomy is a form of automation and its fear is an old one and the reality is automation has worked out for the better.
-
Who cares even if a human is in charge of killing. U cant trust people either.
And if a robot is deciding when to pull a trigger, its a person that programmed it when to, overall.
Also, robots are deadly even without given a gun. Also giving a robot a gun is completely overkill, if you actually code the animation decenty.
-
Playing with words doesn't change a thing.
That is why I posted the response, autonomy is a form of automation and its fear is an old one and the reality is automation has worked out for the better.
No, automation is not very dangerous indeed. But autonomy is not the same as automation. Automation implies supervision, autonomy does not. Our creations are more and more autonomous, which means we have less control. It's exactly like genetic engineering: believing that modified organisms won't spread out in nature is wishful thinking to me. Would you agree?
-
believing that modified organisms won't spread out in nature is wishful thinking to me. Would you agree?
Yes indeed, and they dont care. so we are doomed.
-
No, automation is not very dangerous indeed. But autonomy is not the same as automation. Automation implies supervision, autonomy does not. Our creations are more and more autonomous, which means we have less control. It's exactly like genetic engineering: believing that modified organisms won't spread out in nature is wishful thinking to me. Would you agree?
No, automation does not imply supervision, I don't know where you got that notion from but look up any definition of automation and supervision is not part of it.
-
I'll keep my home offline.
When I was a kid I used to program my Amstrad CPC464. Lot of fun. I don't need a connection to enjoy coding stuff. Maybe, I'll eventually disappear from the grid, and just take care of my garden.
Thats similar to coding an arduino today, I just got the shock of my life with the little state and very slow hz.
I had to chop everything back to byte sized variables, helped alot!!! :2funny:
-
No, automation is not very dangerous indeed. But autonomy is not the same as automation. Automation implies supervision, autonomy does not. Our creations are more and more autonomous, which means we have less control. It's exactly like genetic engineering: believing that modified organisms won't spread out in nature is wishful thinking to me. Would you agree?
No, automation does not imply supervision, I don't know where you got that notion from but look up any definition of automation and supervision is not part of it.
From a dictionary.
Of course it does! An auto-pilot is a tool that human pilots can use during the flight, while an autonomous plane is an unmanned aircraft. Don't tell me you don't understand the difference between a plane with an autopilot, and a drone.
747 and A320 are not drones, not because it's technically impossible, but because they would be commercial unicorn-poop failures.
To me, you're hiding behind words to avoid facing the topic.
-
From a dictionary.
Of course it does! An auto-pilot is a tool that human pilots can use during the flight, while an autonomous plane is an unmanned aircraft. Don't tell me you don't understand the difference between a plane with an autopilot, and a drone.
747 and A320 are not drones, not because it's technically impossible, but because they would be commercial unicorn-poop failures.
To me, you're hiding behind words to avoid facing the topic.
LoL, you're really lost here. You are using a definition of autopilot and NOT the definition of automation and while autopilots are monitored so too are drones. If you had looked up the FAA regulations (https://www.faa.gov/newsroom/small-unmanned-aircraft-systems-uas-regulations-part-107?newsId=22615) you would have known that! Even Amazon's delivery drones are supervised (https://www.cnbc.com/2020/08/31/amazon-prime-now-drone-delivery-fleet-gets-faa-approval.html) since the operator is using the drones camera and a tracking system to monitor it. Even military drones inclusive of cruise missiles are supervised. Both autopilots and drones could fly autonomously.
So look up the term "Automation" not one example of automation...
-
Ah, forget about it, you're missing the point.
:)
-
Ah, forget about it, you're missing the point.
:)
No, you're missing the point, autonomy, with respect to machines, is a form of automation. That you not get that is very puzzling... :o
-
I understand perfectly why you say that autonomy is a form of automation, it is indeed pretty obvious to anyone here why you say that. However, these 2 words carry different meanings, they are not perfect synonyms. You may think that the difference between these words is subtle, but it is very real. Now, beyond words, what matters is the level of autonomy, or level of freedom if you prefer, of the machines, which grows as time passes. I'm not saying it's fundamentally wrong, I'm just saying there's nothing wrong with being careful about it.
Greater autonomy by definition implies less control. Personally I wouldn't get on an unmanned plane, and you can say whatever you want, I think no one in the world would, because that would just be crazy.
:)
You seem (but I may be wrong) to imply that current technological development is comparable to that of previous eras in human history, but that is not my opinion. I think that the current technological evolution imposes on us challenges which humans have never been confronted with, and moreover, on a global scale.
-
You seem (but I may be wrong) to imply that current technological development is comparable to that of previous eras in human history, but that is not my opinion. I think that the current technological evolution imposes on us challenges which humans have never been confronted with, and moreover, on a global scale.
No, my point is not that current technological development is comparable to previous eras of human history, but the fear of this development is comparable to that of previous eras. What you and Elon Musk project is paranoia with the old adage of "We won't need humans anymore", which was the claim of previous eras. How humanity will adopt, implement and create new opportunities with advanced AI is not clearly or completely known and it is comparable to Benjamin Franklin's perspective where he could not have possibly imagined the use and impact of electricity. The paranoid vocalist of AI fail to realize that machines serve humanity and free markets need people...
-
Please don't put me and Musk in the same bag, I don't know his exact opinions. I'm talking for myself.
I could agree on fear being comparable.
But we're facing something that never happened before: something that's not human that, one day, will be more intelligent than humans, that will be able to decide by themselves what they want to become, and what their role in society *should* be. This is comparable to nothing we've invented before. I'm not afraid, I'm just able to see how big this is going to be.
-
Please don't put me and Musk in the same bag, I don't know his exact opinions. I'm talking for myself.
I could agree on fear being comparable.
But we're facing something that never happened before: something that's not human that, one day, will be more intelligent than humans, that will be able to decide by themselves what they want to become, and what their role in society *should* be. This is comparable to nothing we've invented before. I'm not afraid, I'm just able to see how big this is going to be.
Who said that AI will be able to decide what they want to become? I believe I posted a message describing how too much of what AI will be by the media and paranoid vocalists is anthropological. An AGI does not have to implement such social paradigms as "what it wants to become", it's just nonsense. We build machines to serve humanity, that's not going to change with the development of AGI. Realize that even if AI is allowed to improve its self it is in the context of how it can better serve humanity.
As an example: my approach to AGI involves the use of qualia which implements signal states inclusive of emotions to arbitrate choices. This allows for an AGI to choose between the lesser of two evils by evaluating consequences in a context of a quality of life. Now that sounds like selfishness that is serving the self-interest of the AI. Now let's look at how nature uses this kind of selfishness that then transforms into selflessness. Women have a motherly sense, an emotion towards infants that is very gratifying and therefore selfish to the ends of the woman pursuing that effect to care for the infant. However selfish that gratification is it ends up motivating selfless actions that protect and care for the infant. From that perspective, we could do the same for AGI, where it finds an emotional gratification to serve humanity as well as an emotional distress if it is lacking in that regard.
You see we design the AGI to serve us not the other way around. We don't let AGI decide what it wants to become in any context that challenges humanity's power over machines.
-
Who's that "we" you're talking about? does the entire mankind do AGI the frankinstein way? Don't forget you're just one guy on a forum, you're not The One father of AGI. The world doesn't care about your dos and donts and the truth is you don't even know what kind of AGI, if any, might already exist somewhere.
Edit: I think I know now why you don't understand the issue. You approach it as if it was a top-down process, but it's not. We're in a bottom-up / emergence context: if something can happen, it will.
-
Who's that "we" you're talking about? does the entire mankind do AGI the frankinstein way? Don't forget you're just one guy on a forum, you're not The One father of AGI. The world doesn't care about your dos and donts and the truth is you don't even know what kind of AGI, if any, might already exist somewhere.
Edit: I think I know now why you don't understand the issue. You approach it as if it was a top-down process, but it's not. We're in a bottom-up / emergence context: if something can happen, it will.
What part of it's an "example" didn't you get? You create paranoia because quite frankly you don't understand that you're anthropomorphizing. You then go on to create conspiracy theories of "There's some super AGI lurking out there" that will what, figure out how to take over the world! ;D And that bottom-up stuff is laughable I bet you watch those "Terminator" movies over and over again. :2funny:
-
I didn't say that there is a "super AGI lurking out there". Misreading again einstein?
:)
-
I didn't say that there is a "super AGI lurking out there". Misreading again einstein?
:)
I see you suffer from short-term memory loss, must be aging that's catching up with you. So, to quote you literally:
the truth is you don't even know what kind of AGI, if any, might already exist somewhere.
:2funny:
-
Maybe you should take reading lessons einstein.
you don't even know
This means "you don't have the knowledge of".
what kind of AGI
This means "the nature of, the technologies used in, the purpose of, ..., AGI".
if any
This means "if there exists an AGI", which means, again, we don't know whether an AGI exists somewhere.
might already exist somewhere.
I'm tired of explaining things to you. Now I'll stop "talking" with you because well, you're not intelligent enough to deal with the topic.
Talk to yourself.
-
I suspect we are in the superposition of sentence interpretation, maybe ur both right?
-
Maybe the dumb guy should debate on the topic with the racist guy. ;D
Have fun! ;)
-
I take it as a compliment.
-
Apology accepted
-
Whats wrong with that dude? He looks like a legend to me!!! :2funny:
-
I take it as a compliment.
Wouldn't it make more sense to judge people by their personality?
-
I wish I could retort, But I am bullied into silence by the PC patrol.
-
Debates about race are not usually germane to the topic of AI, and I would hate it if gratuitous insults ever drove a non-white AI enthusiast away from this forum. So in this case I support the actions of the "PC patrol." We simply don't need that kind of talk here.
-
Alright. If the consensus is that we won't talk about it, that's fine with me. Lets just all abide by this decision.
-
Please accept my apologies frankinstien and MagnusWootton.
-
Please accept my apologies frankinstien and MagnusWootton.
Apology accepted, I removed the negative comment with the clown pic.